[Alpha] 386.2

[Mutzli]

I'm also one of those unlucky Comcast customers (It's the only game in town). Using cake since its release and never had any issues with it. For me it works much better than any of the other integrated options. I'll have to keep using the script since I need wash and some other Comcast specific options enabled.

Btw. Just notice the updated script version with support for additional models. Thanks @ttgapers and everyone else for keeping this script going.

 

[rlj2]

Getting non-stop "bcm63xx_nand ff801800.nand: intfc status c80000e0" in the logs on a AC86U, cpus stay pegged at 100%. Everything works with 386.2 , might need to factory reset.
EDIT: a reset fixed the log errors. The 100% cpu was unbound stuck in some awk and sed loops. I just removed it.

 

[RMerlin]

Is it properly installed?

It was, I even copied the whole etc folder provided by the iproute2 package when it didn't work with just the map file. If that file is missing, then iproute2 will report the missing map and recommend the value to associate to the selected parameter. If the file is there, then it returns a bogus error message where it doesn't understand the parameter provided to match (something like "what is cmp(blah,blah)")

I'll have another try at some point by reviewing your suggested changes. For now my focus is back on GPL merges.

 

[Make WiFi Great Again]

i understand that cake is not available for the ac68u but i am seeing the option to enable cake under qos on my 68u. infact according to the router i currently have it enabled and running lol so i guess this is a cosmetic error?

 

[AntonK]

Cake isn’t available on the AC68U (kernel too old). Only HND models.

Is there a list of which asus routers are HND models?

 

[John Fitzgerald]

Is there a list of which asus routers are HND models?

I think this is the list that's applicable: https://github.com/ttgapers/cakeqos-merlin#cakeqos-merlin

Maybe not a full list of HND.

 

[AntonK]

I think this is the list that's applicable: https://github.com/ttgapers/cakeqos-merlin#cakeqos-merlin

Maybe not a full list of HND.

Thanks!

 

[brummygit]

i understand that cake is not available for the ac68u but i am seeing the option to enable cake under qos on my 68u. infact according to the router i currently have it enabled and running lol so i guess this is a cosmetic error?

Maybe the placebo effect will reduce your bufferbloat?

Agreed, I guess it's something for RMerlin to address in a future fix.

 

[Make WiFi Great Again]

Maybe the placebo effect will reduce your bufferbloat?

haha the family swears the interwebz is 50% faster!

 

[Tekneek]

haha the family swears the interwebz is 50% faster!

I have put in ridiculous hours trying to fine tune performance and I get "seems the same, maybe worse." I no longer talk about my efforts on that front.

 

[hifiwifi]

haha the family swears the interwebz is 50% faster!

I guess I'm in the same boat here. Video buffering is a lot shorter without any QOS, I guess.

 

[kernol]

Nothing in it for RT-AX86U ... but could help RT-AC88U and RT-AX56U ... change log so far ...

386.2 (xx-xxx-2021)
  - UPDATED: inadyn to 2.8.1.
  - FIXED: Missing Game Mode to RT-AC88U
  - FIXED: Non-working 160 MHz settings for RT-AC88U region/versions
           that should support it
  - FIXED: Missing Instant Guard to RT-AX56U and RT-AC88U.

Lots in it ... since I posted the above response a month ago ...
No glitches so far on both RT-AX86U's - and built in Cake-Qos working purrrfectly !

 

[Kingp1n]

I tested a little of the alpha and cake. Aside, from the VPN leak I was having I switched back to the stable version. Looking really good though. I'll wait til beta release. Keep at the great work all.

 

[Intrepid2007]

I decided to go back one version. I will skip this alpha and wait until the next version is released. Thanks for all the efforts

Hi Eric ( @RMerlin ),

I installed the latest alpha version on my RT-AX88U, which was released March 4, 2021.

After flashing this firmware version I noticed that selective routing doesn't work well anymore. Maybe something got broke in this version (?)


Test case #1:
LAN IP client: 192.168.1.10

Settings for VPN client #1:
Accept DNS Configuration: Exclusive

View attachment 31584
. VPN is running.
. webpages don't load (no DNS!)
. Pinging 1.1.1.1 results in replies.

When 'Accept DNS Configuration' is set to 'strict' or 'relaxed', internet access works but the VPN's DNS server is NOT used. Instead, the WAN's DNS server is used. This is to be expected when the VPN's DNS server is not responding.


Test case #2:
LAN IP client: 192.168.1.10

Settings for VPN client #1:
Accept DNS Configuration: Exclusive

View attachment 31585

Now everything works fine (selective routing disabled). No DNS leaks, DNS server from VPN provider is used.

For the time being I disabled selective routing.

If you want me to test something, please let me know!

Regards,
Alex

 

[skeal]

@RMerlin You said it would be user friendly, wow does cake work well with just what you've built. Great work. New router makes all the difference...lol

 

[Wisiwyg]

I have put in ridiculous hours trying to fine tune performance and I get "seems the same, maybe worse." I no longer talk about my efforts on that front.

Second that... I now get up early and do anything that I'd like to try before 7am, long before anyone else gets up, so I don't get the texts about "what did you do to the internet"?

 

[Wisiwyg]

I decided to go back one version. I will skip this alpha and wait until the next version is released. Thanks for all the efforts

This current and the prior 386.2 alpha1 versions are the best of the 386.x I've run...

 

[RMerlin]

This current and the prior 386.2 alpha1 versions are the best of the 386.x I've run...

Not much difference between 386.1_2 and 386.2 alpha 1, cake being the only major change. The main course will be in alpha 2 which switches to 386_42095, and contains updated SDKs for all HND models among other things.

 

[Torson]

I'm using @Xentrk's selective routing. After installing the latest 386.2 alpha (w/cake) the RPDB rules are lost and selective routing is unusable. Reverted back to 386.1_2 and all is working well. Sorry, did not have time to capture the logs, but hopefully someone will...

EDIT: just realized that something similar was reported earlier in the thread - http://www.snbforums.com/threads/alpha-386-2.70020/post-669817

 

[RMerlin]

@Odkrys Looks like the issue is iproute2's error report is pretty poor. I was testing with the following command, taken from the man page:

tc filter add dev eth0 basic match 'cmp(u16 at 3 layer 2 mask 0xff00 gt 20)'

Which returned:

What is "cmp(u16 at 3 layer 2 mask 0xff00 gt 20)"?

Seemingly indicating it didn't recognize the cmp() match. However, when I provide a flowid:

tc filter add dev eth0 basic match 'cmp(u16 at 3 layer 2 mask 0xff00 gt 20)' flowid 1

now, it properly gets added.

admin@RT-AX86U-5F38:/tmp/home/root# tc filter show dev eth0
filter parent 8001: protocol all pref 49152 basic
filter parent 8001: protocol all pref 49152 basic handle 0x1 flowid :1
  cmp(u16 at 3 layer 2 mask 0xff00 gt 20)

I hate iproute2's syntax... Written by engineers, for engineers.